The invention in general relates to metal oxide materials with optimized stoichiometry in integrated circuits, having high dielectric constants, capacitances and other electrical properties that are virtually independent of voltage and temperature.
It is well known that there is a need for a high dielectric constant material suitable for use as a charge storage medium in integrated circuits and related applications, such as the bypass capacitor in microwave monolithic integrated circuits (MMICs). The most commonly used dielectric material used for charge storage in integrated circuits is silicon dioxide, which has a dielectric constant of about 4. Other common dielectric materials used for charge storage, such as Si3N4, Al2O3 and Ta2O5, also have low dielectric constants, ranging from 4 to 20. Storage capacitors using such a material must have a large area to provide the capacitive values required in state-of-the-art integrated circuits. These large areas make it difficult to reach high densities of capacitive components in an integrated circuit or MMIC. However, the use of other materials to provide the dielectric in integrated circuits has been hindered by the fact that commonly-used materials with high dielectric constants often have undesirable properties. For example, ferroelectric materials such as lead zirconium titanate (PZT) are known to have high dielectric constants and therefore have been proposed as candidates for high dielectric constant memories. However, it is well-known in the art that ferroelectric materials should not be used in high dielectric constant DRAMs since they produce a large switching charge at the coercive voltage, which switching charge would mask the conventional linear charge that the DRAM utilizes as a memory storage medium. See, O. Auciello, J. F. Scott, and R. Ramesh, xe2x80x9cThe Physics of Ferroelectric Memoriesxe2x80x9d, Physics Today, 51, No. 7, July 1998, pp. 22-27, particularly xe2x80x9cBox 1xe2x80x9d on page 24. Moreover, PZT and other high dielectric materials display a significant, nonlinear dependence of dielectric constant on temperature and applied voltage. If used as a capacitor dielectric or a bypass capacitor, the material causes the capacitance value to change with temperature and voltage. Materials in which the electronic properties change with temperature and voltage are generally undesirable in integrated circuits and MMICs, since the effective operation of the circuits requires that the electronic properties have specified values, and integrated circuits should be able to operate over a range of temperatures and voltages. PZT and other high dielectric materials often have electronic properties that do not scale; that is, the properties change significantly as the material is made thinner. Such lack of scaling can place stringent requirements on processing that are difficult to meet, and require wholesale redesign of circuits as they are made more dense. Thus, it would be highly desirable to have a high dielectric constant material for which the electronic properties scale and are essentially independent of temperature and voltage.
It would also be useful to have a high dielectric constant capacitor that can be used generally in integrated circuits, such as for gate dielectric films in metal-oxide-semiconductor field effect transistors (MOSFETs), ferroelectric field effect transistors (ferroelectric FETs), and other transistors, as buffer layers to prevent unwanted interactions between certain materials, as diffusion barriers to prevent diffusion of elements from one layer to another, and as interlayer dielectrics. In each of these applications it is useful that the material have electronic values that do not change with temperature and voltage. In addition, each of these applications have additional electronic requirements. It is particularly difficult to predict if a high dielectric material will be effective in gate insulator films in MOSFETs, ferroelectric FETs, and other transistors because, in this application, it must meet multiple, often conflicting, requirements. In addition to having a capacitance that is flat with respect to temperature, voltage and thickness variations, they must have low leakage current and a high breakdown voltage, they must not alter the threshold voltage of the transistor as a function of gate voltage, and they must be an effective barrier against charge injection. Buffer layers must be compatible with both the material they buffer and the surrounding integrated circuit materials. Diffusion barriers must be effective in preventing migration of particular elements at elevated temperatures and must not themselves include elements that can migrate. Interlayer dielectrics must have low leakage currents and have a high breakdown voltage. As integrated circuits become smaller, all the above requirements become more and more stringent, since thicknesses of the materials decrease and the distances between dissimilar materials shrink. The paucity of materials that have a dielectric constant greater than 20, but do not possess problematic properties, is considered to be one of the serious roadblocks to higher density integrated circuit memories.
Recently, commercial ferroelectric memories, particularly ferroelectric random access memories (FERAMs), have become available. While practical FERAMs have not yet reached the densities of DRAMs, rapid advances are being made in this area that suggest that these memories may soon become competitive with DRAMs. Since these memories, like DRAMs, are optimally intended for use in environments in which the temperature and voltage can change substantially, it would be highly useful to have ferroelectric materials for which the electronic properties scale and are essentially independent of temperature and voltage.
The present invention solves the above problem by providing an integrated circuit incorporating a thin film of metal oxide for which the electronic properties have low dependence on voltage and temperature. The invention also provides a liquid precursor for forming a thin film of metal oxide for which the electronic properties scale and have low dependence on voltage and temperature. In addition, the invention provides a method for applying a liquid precursor to an integrated circuit substrate and treating the applied precursor to form the thin film of the metal oxide for which the electronic properties have low dependence on voltage and temperature. Preferably, the metal oxide is a high dielectric constant material, and most preferably it is not ferroelectric. However, some of the materials are ferroelectric, and therefore will be useful in ferroelectric devices, such as FERAMs.
An embodiment of the invention is an integrated circuit comprising a high dielectric constant metal oxide insulator thin film or ferroelectric metal oxide thin film, wherein the metal oxide is selected from the group consisting of tungsten-bronze-type oxides, pyrochlore-type oxides, and combinations of an interlayer oxide with an oxide selected from the group consisting of pyrochlore-type oxides and tungsten-bronze-type oxides. Typically, the interlayer oxide is Bi2O3.
In an embodiment of the invention, the metal oxide has a stoichiometry represented by a formula selected from the group consisting of AB2O6, A2B2O7 and A2Bi2B2O10, wherein A represents A-site atoms selected from the group of metals consisting of Ba, Bi, Sr, Pb, Ca, K, Na and La; and B represents B-site atoms selected from the group of metals consisting of Ti, Zr, Ta, Hf, Mo, W and Nb.
An integrated circuit according to the invention contains a thin film of metal oxide with a thickness preferably ranging from 1 nanometer (nm) to 500 nm. Metal oxide material according to the invention is preferably non-ferroelectric and has a relatively high dielectric constant, i.e., a dielectric constant value, xcex520, of 20 or higher. It shows negligible dependence of capacitance on temperature and on external applied voltage, having low Vcc values and low Tcc values. Typical applications of these materials are as the dielectric material in the capacitor of DRAMs, the gate dielectric of transistors, including ferroelectric FETs, as buffer layers to prevent unwanted interactions between certain materials, particularly layered superlattice materials and semiconductors, as diffusion barriers to prevent diffusion of elements from one layer to another, and as interlayer dielectrics. Some of the materials according to the invention are ferroelectric, and a typical application of these materials is as part of a memory cell in an integrated circuit ferroelectric random access memory (FERAM) or as a ferroelectric gate in ferroelectric FETs.
A liquid precursor according to an embodiment of the invention comprises metal atoms in relative molar proportions for forming a metal oxide selected from the group consisting of tungsten-bronze-type oxides, pyrochlore-type oxides, and combinations of an interlayer oxide, such as Bi2O3, with an oxide selected from the group consisting of pyrochlore-type oxides and tungsten-bronze-type oxides. In an embodiment of the invention, the liquid precursor contains metal atoms having a stoichiometry represented by a formula selected from the group consisting of AB2O6, A2B2O7 and A2Bi2B2O10, wherein A represents A-site atoms selected from the group of metals consisting of Ba, Bi, Sr, Pb, Ca, K, Na and La; and B represents B-site atoms selected from the group of metals consisting of Ti, Zr, Ta, Hf, Mo, W and Nb.
An embodiment of the method of the invention comprises steps of: providing a substrate; providing a liquid precursor for forming a thin film of the above-discussed metal oxide; applying the liquid precursor to the substrate to form a liquid coating on the substrate; and treating the liquid coating to form a solid thin film of the desired metal oxide. The step of applying may comprise spin-coating a liquid coating of the precursor on the substrate, misted deposition, dip coating or other liquid application process. The step of treating may comprise a process selected from the group including: exposing to vacuum, exposing to ultraviolet radiation, drying, heating, baking, rapid thermal processing, and annealing. An embodiment of the inventive method includes providing a liquid precursor comprising metal atoms in relative molar proportions corresponding to an interlayer oxide, such as Bi2O3, added to a precursor of a perovskite or a pyrochlore-type oxide compound.
The excellent properties of the material and the ability to make very thin films of the material that retain those properties over wide ranges of temperature and voltage make the invention of particular importance for integrated circuits, but it is also important for any type of charge storage device. Numerous other features, objects and advantages of the invention will become apparent from the following description when read in conjunction with the accompanying drawings.